精度、召回率、精度和 f 测量的小数部分

Question

我正在进行一项 CNN 实验，并使用准确率、召回率、精度和 f-measure 指标评估构建的分类器。

结果值只有两个小数位（即 0.95）。我怎样才能有至少 4 个小数位（即 0.9532）

此外，我如何将结果值四舍五入以获得四舍五入的数字的下限或上限？

from keras.models import Sequential
from keras.layers import Conv2D,Activation,MaxPooling2D,Dense,Flatten,Dropout
import numpy as np
from keras.preprocessing.image import ImageDataGenerator
from IPython.display import display
import matplotlib.pyplot as plt
from PIL import Image
from sklearn.metrics import classification_report, confusion_matrix
import keras
from keras.layers import BatchNormalization
from keras.optimizers import Adam
import pickle
from keras.models import load_model
classifier = load_model('32_With_Dropout_rl_001_1_layer_2.h5')
classifier1 = load_model('32_With_Dropout_rl_001_2_layers_2.h5')
classifier2 = load_model('32_With_Dropout_rl_001_3_layers_2.h5')
test_datagen = ImageDataGenerator(rescale = 1./255)
batchsize=10
test_set = test_datagen.flow_from_directory('/home/osboxes/Downloads/Downloads/Journal_Paper/Benign_Malicious/Spectrogram/Test/',
                                           target_size = (200,200),
                                           batch_size = batchsize,
                       shuffle=False,
                                           class_mode ='categorical')
Y_pred = classifier.predict_generator(test_set, steps= 1023 // batchsize+1)
y_pred = np.argmax(Y_pred, axis=1)
print('Confusion Matrix 1 layer')
print(confusion_matrix(test_set.classes, y_pred))
print('Classification Report')
target_names = test_set.classes
class_labels = list(test_set.class_indices.keys()) 
target_names = ['Bening', 'Malicious'] 
report = classification_report(test_set.classes, y_pred, target_names=class_labels)
print(report) 

Y_pred = classifier1.predict_generator(test_set, steps= 1023 // batchsize+1)
y_pred = np.argmax(Y_pred, axis=1)
print('Confusion Matrix 2 layers')
print(confusion_matrix(test_set.classes, y_pred))
print('Classification Report')
target_names = test_set.classes
class_labels = list(test_set.class_indices.keys()) 
target_names = ['Bening', 'Malicious'] 
report = classification_report(test_set.classes, y_pred, target_names=class_labels)
print(report) 

Y_pred = classifier2.predict_generator(test_set, steps= 1023 // batchsize+1)
y_pred = np.argmax(Y_pred, axis=1)
print('Confusion Matrix 3 layers')
print(confusion_matrix(test_set.classes, y_pred))
print('Classification Report')
target_names = test_set.classes
class_labels = list(test_set.class_indices.keys()) 
target_names = ['Bening', 'Malicious'] 
report = classification_report(test_set.classes, y_pred, target_names=class_labels)
print(report) 


#f = open('32_With_Dropout_rl_001_1_layer', 'rb')
#history = pickle.load(f)

#f = open('32_With_Dropout_rl_001_2_layers', 'rb')
#history1 = pickle.load(f)


#f = open('32_With_Dropout_rl_001_3_layers', 'rb')
#history2 = pickle.load(f)

Answer 1

在计算分类报告时，您只需使用output_dict参数。如果设置为True，则输出作为DICT（而不是字符串）返回。然后，您可以访问所有字段并根据需要修改它们（例如使用@ 987654324或@ 487654325 @）。

from math import ceil, floor

# do your stuff here
report = classification_report(..., output_dict=True)
# use the values of the report dictionary as you wish
print(floor(report[...]))

如果要将报告作为字符串作为字符串获取，则应使用digits参数：

report = classification_report(..., digits=4)

检查文档here。